financial differences and similarities of islamic … · similarities between banks in terms of...

Journal of Business, Economics & Finance (2015), Vol4 (2) Yildirim, 2015

232

FINANCIAL DIFFERENCES AND SIMILARITIES OF ISLAMIC BANKS: A STUDY ON QISMUT COUNTRIES

DOI: 10.17261/Pressacademia.2015211617

İsmail YILDIRIM1

1Hitit University. E-mail: [email protected]

Keywords: QISMUT Countries, Islamic Banks, Cluster Analysis, Multidimensional Scaling Analysis Jel Classification: G21; G15

ABSTRACT

Six countries, namely Qatar, Indonesia, Saudi Arabia, Malaysia,

UAE and Turkey (abbreviated as QISMUT) are considered to play

an important role in the future international development of the

Islamic finance. QISMUT countries have a majority Muslim

population who prefers Islamic finance for their banking needs.

QISMUT countries account for 38 million customers which is 67%

of the global Islamic bank customer base.

The focus of this study is to define similar (or different)

international banks using the financial data collected from Islamic

banks operating in QISMUT countries between 2012 and 2014. In

this context, 56 Islamic banks were categorized in the light of

their financial ratios for the period between 2012 and 2014 using

cluster analysis. There was also an attempt to identify these

clusters using multidimensional scaling analysis. Positions of

QISMUT Islamic banks in the multidimensional space were

defined using multidimensional scaling analysis. Results showed

that Islamic banks operating in QISMUT countries are similar in

terms of their financial structure.

1. INTRODUCTION

Islamic banks have been making efforts to increase their productivity and to improve their performance in order to gain sustainability as a result of the recent globalization movement (Mghaieth and Mehdi, 2014). Recent global mortgage crisis brought Islamic finance forward as an alternative in terms of investment and banking (Smola and Mirakhor, 2010). Islamic banking has become an integral part of the global finance structure particularly with its immunity to the recent banking and financial crises (Aldohni, 2015).

Islamic banking has gained momentum and acknowledgement especially in the Middle-East and Southern Asia when compared to the rest of the world (Ariff, 2014). Islamic banking is widely preferred in these regions as the Muslim population is in the majority and the Islamic rules are internalized. These regions with their concentrated Muslim population are leading the Islamic finance. QISMUT is a composition of countries which will make its mark in the future of Islamic finance and banking in both regional and global terms.

Year: 2015 Volume: 4 Issue: 2

Journal of Business, Economics & Finance ISSN: 2146 - 7943

mailto:[email protected]


233

Six rapid growth markets (QISMUT countries: Qatar, Indonesia, Saudi Arabia, Malaysia, UAE and Turkey) will play an important role in the globalization of Islamic banking industry. Having a wide pool of intellectual capital and funds, these countries may well be the drivers of a growth wave in current and new markets. Two third of the 38 million Islamic bank customers in the world reside in QISMUT countries. A Compound Annual Growth Rate (CAGR) of 19.7% is expected for the total assets of Islamic banks operating in QISMUT countries between 2013 and 2018 adding up to USD $1.6 trillion (2012: USD $567 billion) (Nazim, Bellens, 2014). It is estimated that total assets of the Islamic banks in QISMUT countries reached $662 billion by 2013. Global Islamic banking assets, on the other hand, are anticipated to reach $3.4 trillion by 2018 (Lackmann, 2014).

This study aims to investigate if the QISMUT Islamic banks which are anticipated to improve their profits in a global scale in the years to come have similarities in terms of their financials. QISMUT country classification is a rather new one. Previous research involves studies performed in order to include countries in this group such as research on the Middle-Eastern (including Gulf Cooperation Council countries) countries, North African (MEAN) countries, Gulf Cooperation Council (GCC) countries, and Organization of the Islamic Conference (OIC) countries. These classifications were made mainly through taking geographic, economic, and political aspects into consideration. As a new classification and expected to become the leader of Islamic banking in the close future, do QISMUT Islamic banks show similarities with respect to their lucrativeness, growth, risks, and DuPont ratios? Establishing these similarities will have notable implications for QISMUT classification in the future. This study is significant in this context.

This study uses 15 financial variables obtained from annual financial statements and statements of receipts published by Islamic banks operating in the QISMUT countries. It was attempted to identify the similarities and discrepancies between Islamic banks using variables such as Lucrativeness, Growth, Risks, and DuPont Ratios with Cluster Analysis and Multidimensional Scaling Analysis.

2. LITERATURE REVIEW

Literature includes Cluster Analysis and Multidimensional Scaling Analysis in the identification of similarities and discrepancies between various objects. Lucotte (2015), Knotek (2014), Sorensen and Gutierrez (2006) classified the banks operating in Eurozone using a Cluster Analysis. Lucotte (2015) also used Cluster Analysis in order to be able to classify banks operating in Eurozone in the aftermath of the recent global financial crisis. Knotek (2014) tested the similarities and discrepancies between banks operating in Eurozone using a Cluster Analysis. Sorensen and Gutierrez (2006) made use of Cluster Analysis techniques in order to identify the uniformity of the banks operating in Eurozone between 1998 and 2004. In their study, Western (Germany, France, Belgium and to some extent the Netherlands, Austria and Italy) and Central European countries were clustered with Spain, Portugal, and Greece. Irish and Finnish banks were clustered in a separate group however. The study concluded that there is a convergence between them and the banks operating in Western and Central Europe. Kowal, Yeleyko, and Kharkhalis (2014) analyzed the credit operations of foreign-invested banks operating in Ukraine using a hierarchic cluster analysis. Vagizova, Lurie, and Ivasiv (2014) grouped Russian banks using Cluster Analysis.


234

Dardac and Boitan (2009) made use of Cluster Analysis in order to examine the profitability and risk ratio of Romanian banks between 2004 and 2006. They have found similarities between banks in terms of risk and profitability. Grzegorz and Dawid (2009) investigated 48 Polish banks classifying them into several groups using Cluster Analysis for a period between 1999 and 2005. Molinero and Cinca (2001) used Multidimensional Scaling Analysis in order to measure the financial failure of banks. Their study investigated 66 Spanish banks for financial stability using Multidimensional Scaling Analysis.

This study has differences from the previous studies on Islamic banks in terms of its scope. Literature consists of studies based on a single country and multiple countries. Most commonly, among these countries are Middle-Eastern (including Gulf Cooperation Council countries) countries, North African (MEAN) countries, Gulf Cooperation Council (GCC) countries, and Organization of the Islamic Conference (OIC) countries. The literature still lacks studies based on QISMUT countries. The reason behind this is that the QISMUT country classification is a rather new one in terms of Islamic banking. This study contributes significantly to the literature.

3. METHODOLOGY AND DATA

This study attempts to identify the financial similarities and discrepancies between Islamic banks using Cluster Analysis and Multidimensional Scaling Analysis.

Cluster analysis is a multivariate statistics method which aims to sort observations set into a limited number of groups or clusters. This kind of a sorting takes place when the observations obtained from the same group are similar while they are different from the observations obtained from other groups (Neil, 2002). Cluster Analysis is a commonly used method among non-structural multivariate analysis methods. Methodology of cluster analysis consists of algorithms which arrange a given dataset into subsets (Izenman, 2008).

Cluster analysis makes use of hierarchical and non-hierarchical clustering method in order to sort units into appropriate groups (Hand et al., 2001). Hierarchical cluster analysis is a significant statistical method used in order to identify uniform clusters. This method makes use of differences of distances between objects during the formation of clusters. Results are then shown in a hierarchical tree diagram.

However, clustering method offers fast measurements, hierarchical methods are not convenient for the analyses of larger samples (Hair, 1998). K-Means technique will be explained as it is the technique used in this study.

3.1. Non-Hierarchical Clustering (K-Means) Method

Non-hierarchical clustering analysis is a method designed to cluster units when the number of clusters is fixed to K. This method includes two widely preferred techniques. One being the K-Means technique and the other is the probability technique. K-Means is one of the most commonly used non-hierarchical clustering techniques (Lin and Chen 2006). K-Means technique involves “d” number of variables of the dataset, “x”, and N number of feature vectors which are able to cluster and classify into K number of clusters (Burn and Goel, 2000). This method starts with the identification of K number of clusters. Each object in the dataset is allocated in a cluster (Rao and Srivinas, 2006).


235

A possible change in the results must be considered as they will depend on the K-means centers defined in the beginning. Algorithm selects a number of k-means centers in order to reach a better solution every time it runs. It attempts to find the smallest value of the objective function while running these selections (Liu and Yu, 2009).

When each x1,x2,………,xn value is taken as observation vectors with a number of “d” variables and a point in the multidimensional N space and a1n,……..,akn value was selected as the cluster center for each group observation;

𝑊𝑛 =1

𝑛 ∑ 𝑚𝑖𝑛𝑛

𝑖=1 |𝑥𝑖 − 𝑎𝑗ℎ|2

Using this formula, observations are allocated to the closest cluster. When n is the number of data and p is the quality factor in the data matrix;

Disparity Matrix

the distance between two data is d(i,j). When q=2 and d is Euclidean Distance, the similarity and disparity between data is then measured with;

Features:

d(i,j) ≥ 0

d(i,j) = 0

d(i,j) = d(j,i)

d(i,j) ≤ d(i,h)+d(h,j)


236

Cluster pattern recognition and image processing is widely used in many fields such as economics (particularly in market research), document classification on worldwide web, discovering similar peer groups, data mining, statistics, biology and machine learning (Han and Kamber, 2006).

3.2. Multidimensional Scaling Analysis

Multidimensional Scaling Analysis is a technique which visualizes the “similarity” features between objects with statistical calculations (Machado and Mata, 2013). Distances between objects are shown on a map-like structure (Machado and Mata, 2015). Multidimensional Scaling Analysis aims to graphically display the dataset combined of “n” units on “a” Euclidean space.

Multidimensional Scaling Analysis makes use of stress values in order to determine if the number of dimensions used in the graphical setup is acceptable. Table 1 shows the stress values and quality of compliance obtained from the Multidimensional Scaling Analysis.

Table 1: Stress Values and Quality of Compliance

Stress Values (Compliance) Quality of Compliance

≥ 0,20 Poor Compliance

0,10 < 0,20 Medium Compliance

0,05 < 0,10 Good Compliance

0,025 < 0,05 Very Good Compliance

0,000 < 0,025 Perfect Compliance

Stress values closer to 0 constitute a favorable condition. Perfect or poor compliance is derived from stress values obtained in a range between 0.00 and 0.20. Multidimensional scaling analysis also applies R

2 values for the identification of the quality of compliance. R

2

values above a minimum of 0.60 are indicative of acceptable quality of compliance. Values even higher are indicative of close to perfect quality of compliance (Hair et al., 2006: 654).

This study investigates the Islamic banks operating in the QISMUT countries. According to the ‘World Islamic Banking Competition Report 2013-14’ prepared by Ernst&Young Turkey, an Audit and Counseling Corporation, QISMUT countries (Qatar, Indonesia, Saudi Arabia, Malaysia, UAE and Turkey) which include also Turkey will play an important role in the globalization of Islamic banking industry.

In this respect, the scope of this research is selected to include QISMUT countries in order to be able to reveal the financial disparities and similarities of Islamic banks and to contribute to the literature with valuable outputs. Profiles of these countries with respect to other QISMUT countries are given in Table 2.


237

Table 2: General Profiles of QISMUT Countries (2013)

Population (million)

Islamic Finance Assets

Share in the QISMUT countries

Global Share of the Islamic

Banking Assets

Qatar 2.2 $54,4 billion 9% 3%

Indonesia 248.5 $21,4 billion 4% 1%

Saudi Arabia 31 $245 billion 43% 16%

Malaysia 29.8 $140 billion 22% 8%

UAE 9.3 $83 billion 15% 5%

Turkey 76.1 $50 billion 7% 2%

Total 396.9 $593.8 billion 100% 35%

Source: Ernst and Young; World Islamic Banking Competitiveness Report 2013–14

Muslim population residing in the QISMUT countries constitutes the potential clientele of Islamic banks. The total population of these six countries adds up to 400 million people. Indonesia is the most crowded QISMUT country. Qatar, on the other hand, is the one with the lowest population density. However, Islamic banking assets in Indonesia are the lowest of all proportionally. This share is slightly higher in countries with strict Islamic rules in effect. QISMUT member Saudi Arabia has a 16% market share in the global Islamic banking market. QISMUT member Indonesia, on the other hand, has the smallest market share with 1%. The total share of these 6 QISMUT countries in the global Islamic banking market is 35%. Saudi Arabia is the leading country in terms of Islamic financial assets with 43% when compared to other QISMUT countries. Indonesia, on the other hand, accounts for the lowest share in this respect (see Table 2). Names of the banks included in the analysis are given in Table 3.


238

Table 3: Banks which are Included in the Analysis

Countries Number of bank

Banks included in the analysis

Katar (Qatar)

6

(Q1) Al Khaliji Commercial Bank and Islmaic Banking, (Q2) Commercial Bank, (Q3) Masraf Al Rayan–Qatar, (Q4) Qatar Islamic bank, (Q5) Qatar National Bank, (Q6) Qatar International Islamic Bank

Endonezya Indonesia

15

(I1)Bank Pembangunan Daerah (BPD) Banda Aceh, (I2)PT Bank Tabungan Pensiunan Nasional (BTPN), (I3)PT Bank Danamon, (I4)PT Bank Permata, (I5)PT Bank Sinarmas, (I6)PT Bank Syariah BNI, (I7)PT Bank Syariah BRI, (I8)PT Bank Syariah Bukopin, (I9)PT Bank Syariah Jabardan Banten (BJB), (I10)PT Bank Syariah Mandiri, (I11)PT Bank Syariah Mega Indonesia, (I12)PT Bank Syariah Panin, (I13)PT Bank Syariah Victoria, (I14)PT CIMB Niaga, (I15)PT OCBC NISP

Sudi Arabistan (Saudi Arabia)

11

(S1)Al Rajhi Banking & Investment Corporation, (S2)Alinma Bank, (S3)Arab National Bank, (S4)Bank AlJazira, (S5)Bank Albilad, (S6)Banque Saudi Fransi Islamic Banking, (S7)Riyad Bank, (S8)Samba Financial Group, (S9)Saudi Hollandi Bank, (S10)The National Commercial Bank, (S11)The Saudi Investment Bank

Malezya (Malaysia)

13

(M1)Al Rajhi Banking & Inv. Corp. (Malaysia), (M2)Alliance Islamic Bank Berhad, (M3)Asian Finance Bank, (M4)Bank Islam Malaysia Berhad, (M5)Bank Kerjasama Rakyat Malaysia Berhad, (M6)Bank Pembangunan Malaysia Berhad, (M7)CIMB Islamic Bank Berhad, (M8)Hong Leong Islamic Bank Berhad, (M9)Malayan Banking Berhad, (M10)Maybank Islamic Berhad, (M11)OCBC Al-Amin Bank Berhad, (M12)Public Islamic Bank Berhad, (M13)RHB Islamic Bank Berhad

Birleşik Arap Emirlikleri

(UAE) 7

(U1)Abu Dhabi Islamic Bank, (U2)Emirates Islamic Bank, (U3)Sharjah Islamic Bank, (U4)Ajman Bank, (U5)Siraj Islamic Banking, (U6)Emirates NBD PJSC, (U7)Union National Bank

Türkiye Turkey

4 (T1)Albaraka Turk Participation Bank, (T2)Asya Finance Participation Bank, (T3)Türkiye Finance Participation Bank, (T4)Kuwait Turk Participation Bank

Cluster and Multidimensional Scaling analyses involved a total of 56 Islamic banks; 6 operating in Qatar, 15 in Indonesia, 11 in Saudi Arabia, 13 in Malaysia, 7 in UAE, and 4 in Turkey. Banks included in the analysis are given an individual code. Banks will be represented with these codes in the clusters obtained from the analysis (Table 3).

Financial ratios were used in order to identify the financial similarities and disparities between Islamic banks operating in the QISMUT countries by means of Cluster and Multidimensional Scaling Analysis. Financial ratios included in the analysis are given in Table 4.


239

Table 4: Financial ratios which are Included in the Analysis

Financial Ratio Grouping Code Financial Ratio

Profitability

P1 Efficiency Ratio

P2 Operating Leverage

P3 Non-interest Income / Operation Income

P4 Fee Revenue (%)

P5 Loan Growth (%)

P6 Deposit Growth (%)

Risk R1 Loan Loss Provision (% of Avg. Loans)

R2 Nonperforming Loans (% of Total Loans)

Other Ratios

O1 Tier 1 Risk-Adjusted Capital Ratio

O2 EOP Loans / EOP Deposits

O3 Securities % Avg. Earning Assets

DuPont/Earning Power

D1 Pretax ROA

D2 x Leverage (Assets/Equity)

D3 Pretax ROE

D4 Reinvestment Rate

Fifteen financial ratios and variables were used under 4 categories in order to analyze the financial similarities and disparities or the banks operating in QISMUT countries. Analysis includes efficiency ratio, operating leverage, non-interest income/operation income, fee revenue, loan growth, and deposit growth as profitability ratios/variables. Loan loss provision and nonperforming loans/total loans are taken as risk variables. Tier 1 risk-adjusted capital ratio, EOP loans/EOP deposits, and securities earning ratio are included in the analysis as other ratios. Finally, pretax ROA, assets/equity, pretax ROE, and reinvestment rate are used as DuPont ratios in order to classify the banks in question.

4. FINDINGS

This study aims to identify the financial similarities and outputs of the Islamic banks operating in QISMUT countries and the data collected in this study was investigated using cluster analysis and multidimensional analysis. The analysis results obtained using financial ratios from each bank operating in these 6 countries were then examined for their similarities and disparities. Results from the Cluster Analysis and Multidimensional Scaling Analysis were evaluated separately in order to provide separate results for the analysis periods of 2012-2013-2014.

4.1. Results of the Cluster Analysis

Correct classification rate which is obtained using Discriminant analysis, average shadow stats and Dunn’s partition coefficient were used in order to test the number of clusters generated according to the banking data obtained from 56 Islamic banks operating in QISMUT countries and the relevant cluster validity. Cluster validity results are shown in Table 5 for a reliable analysis.


240

Table 5: Number of Clusters and Cluster Validity Results

# of Clusters

Correct Classification

Rate (Discriminant)

Average Shadow Stats

Dunn’s Partition

Coefficient

2 86% 0.454 0.781

3 92% 0.596 0.795

4 78% 0.454 0.694

5 75% 0.452 0.645

6 67% 0.324 0.620

7 64% 0.358 0.609

8 60% 0.324 0.581

Maximum values of correct classification rate, average shadow stats and Dunn’s partition coefficient are considered in order to identify the number of clusters and cluster validity. According to the correct classification rate findings shown in Table 4, it was found that the best cluster structure includes 3 clusters with a rate of 92%. Similarly, average shadow stats value (0.596) and Dunn’s partition coefficient (0.795) confirm the best cluster structure to include 3 clusters.

K-Means technique, one of the methods of non-hierarchical clustering, was used in the analysis. This technique pre-estimates the number of clusters. According to the analyses, 3 clusters were identified and the distributions of QISMUT Islamic banks by the years are given in Tables 6, 7 and 8.

Table 6: QISMUT Islamic Banks Distribution in Clusters for 2012

Cluster Number

# of Banks in

the Cluster

% of Banks in the

Cluster

Banks in the Cluster

1st

Cluster

41 73% Q1,Q2,Q3,Q4,Q5,Q6,I1,I2,I3,I6,I7,I9,I10,I13I14,I15,S1, S2,S3,S4,S5,S6,S7,S8,S9,S10,S11,M1,M3,M5,M6,U1,U3, U4,U5,U6,U7, T1,T2, T3, T4

2nd

Cluster

1 2% I12

3rd

Cluster

14 25% I4,I5,I8,I11,M2,M4,M7,M8,M9,M10,M11,M12,M13,U2

According to the Cluster Analysis results for 2012 (Table 6), 1st

cluster includes 41 out of 56 banks in the sample. Thus, 1

st cluster accounts for the 73% of all the banks. All 6 Islamic

banks operating in Qatar, all 11 Islamic banks operating in Saudi Arabia, and all 4 Islamic banks operating in Turkey fall in this cluster. The 2

nd cluster accounts for only one bank.

There is only 2% of all the banks included in the analyses fall in this cluster. PT Bank Syariah Panin, a bank operating in Indonesia, is financially different than the other banks included in the sample. The 3

rd cluster includes 4 out of 15 Indonesian Islamic banks, 9 out

of 13 Malaysian Islamic banks and 1 out of 7 UAE banks. There is 25% of all the banks included in the analyses falling in the 3

rd cluster.


241

According to the cluster analysis results for 2012, Islamic banks operating in Qatar, Saudi Arabia, and Turkey have similar properties in terms of profitability, growth and risk.


Cluster Number

# of Banks in

the Cluster

% of Banks in

the Cluster


1st

Cluster

1 2% I12

2nd

Cluster

4 7% M7,M10,M11,M14

3rd

Cluster

51 91%

Q1,Q2,Q3,Q4,Q5,Q6,I1,I2,I3,I4,I5,I6,I7,I8,I9,I10,I11, I13,I14,I15,S1,S2,S3,S4,S5,S6,S7,S8,S9,S10,S11,M1, M2,M3,M4,M5,M6,M8,M9,M12,U1,U2,U3,U4,U5,U6, U7,T1,T2,T3,T4

When the clusters were investigated for the financial variables of 2013 (Table 7), the 1st

cluster includes 1 bank (2% of total); the 2

nd cluster includes 4 banks (7% of total), and the

3rd

cluster includes 51 banks (91% of total). All 6 Islamic banks operating in Qatar, all 11 Islamic banks operating in Saudi Arabia, all 4 Islamic banks operate in Turkey, and all 7 banks operating in UAE fall in the 3

rd cluster. All 4 banks included in the 2

nd cluster operate

in Malaysia. These 4 banks are financially different when compared to the Islamic banks operating in other countries. The 1

st cluster accounts for only one bank, PT Bank Syariah

Panin from Indonesia. As it was the case for 2012, PT Bank Syariah Panin, a bank operating in Indonesia, was financially different when compared to the other banks included in the sample also in 2013. Islamic banks operating in Qatar, Saudi Arabia, UAE and Turkey have similar properties in terms of profitability, growth and risk in 2013.


Cluster Number

# of Banks in

the Cluster

% of Banks in the

Cluster


1st

Cluster

19 34% I1,I4,I8,I9,I11,I13,S4,S10,M1,M2,M4,M8,M9,M12,U2, U4,T1, T2,T3

2nd

Cluster

33 59% Q1,Q2,Q3,Q4,Q5,Q6,I2,I3,I5,I6,I7,I10,I12,I14,I15,S1,S2, S3,S5,S6, S7,S8,S9,S11,M3,M5,M6,U1,U3,U5,U6,U7,T4

3rd

Cluster

4 7% M7,M10,M11,M13

Banks are found to be distinctly differentiated from each other in 2014. The 1st

cluster includes 19 Islamic banks (34% of total), the 2

nd cluster includes 33 Islamic banks (59% of

total) and the 3rd

cluster includes 4 Islamic banks (7%). All 4 banks included in the 3rd

cluster operate in Malaysia. All banks included in the 2

nd cluster along with all bank

operating in Qatar and Saudi Arabia are found to have financial similarities. The 3 Islamic banks operating in Turkey fall in the 1

st cluster which indicates a similarity with Islamic


242

banks operating in Indonesia, Malaysia, and UAE. Islamic banks operating in Qatar and Saudi Arabia have similar properties in terms of profitability, growth and risk in 2014.

Thus the financial similarities and disparities of 56 Islamic banks operating in the QISMUT countries were identified using the cluster analysis. It is important to confirm the classifications obtained using the cluster analysis with Multidimensional Scaling Analysis.

4.2. Multidimensional Scaling Analysis Results

Multidimensional Scaling Analysis makes use of a measure called stress values in order to determine if the number of dimensions used in the graphical setup is acceptable. Table 9 shows the stress values and RSQ values obtained as part of this analysis.

Table 9: Stress and RSQ Results of Multidimensional Scaling Analysis

Analysis Years Stress Values RSQ

2012 0.05963 .99560

2013 0.05836 .99586

2014 0.06967 .99300

Stress values were obtained using Kruskal’s Stress Formula 1. The 2 dimensions were used for Multidimensional Scaling Analysis. The 4 iterations were conducted until the point the value of stress statistics is smaller than 0.001 for both of these dimensions. Accordingly, stress value for 2012, 0.05963 (0.05 < 0.10), is consistent. Stress values of these 2 dimensions account for 99.560%. Stress value for 2013, 0.05836 (0.05 < 0.10), is consistent. Stress values of these dimensions account for 99.586% in 2013. Stress values for 2014 were found to be 0.06967. This value is an indication of good consistency as it is in the range between Stress values of these dimensions account for 99.300% in 2014. Stress coefficients show that the number of dimensions used in the graphical setup obtained from the Multidimensional Scaling Analysis is acceptable for all three years. RSQ values prove to be quite high for each year.

Figure 1 shows the Euclidean Distance Constellation Diagram for the Multidimensional Scaling Analysis of Islamic Banks operating in QISMUT countries in 2012 while Figure 2 shows the Euclidean Distance Model for the same year.

Figure 1: Euclidean Distance Constellation Diagram for 2012


243

As seen in Figure 1, the distances and disparities obtained from the Multidimensional Scaling Analysis of Islamic banks operating in QISMUT countries are aligned in a linear manner (bottom left to top right). In this context, the purpose of the Multidimensional Scaling Analysis was to measure the similarities between the estimated distances generated according to the disparity matrix and the distance matrix obtained through the use of the direct data. According to the results of this research, it was found that the estimated distances were consistent with the actual values as the estimated values were similar to the ones in the distance matrix obtained from data.

Figure 2: Euclidean Distance Model for 2012

According to the Figure 2, 53 out of 56 banks included in the analysis have financial similarities (accumulated around Euclid) as per the Euclidean distance model which shows the relations between units (QISMUT Islamic banks) on a 2-dimensional space. According to the Euclidean distance model for 2012, 3 banks (I12: PT Bank Syariah Panin, M6: Bank Pembangunan Malaysia Berhad and S2: Alinma Bank) have financial disparities when compared to the other Islamic banks.



244


It was found that the distances identified using Multidimensional Scaling Analysis of the QISMUT countries and their disparities have a linear relationship. According to the results of Euclidean distance model for 2013, it was found that the estimated distances were consistent with the actual values as the estimated values were similar to the ones in the distance matrix obtained from data.


According to the Figure 4, 53 out of 56 banks included in the analysis have financial similarities (accumulated around Euclid) as per the Euclidean distance model which shows the relations between units (QISMUT Islamic banks) on a 2-dimensional space. According to the Euclidean distance model for 2013, 3 banks (I12: PT Bank Syariah Panin, M6: Bank Pembangunan Malaysia Berhad and S2: Alinma Bank) have financial disparities when compared to the other Islamic banks. Therefore, Islamic banks operating in Qatar, UAE and Turkey have similar properties in terms of profitability, growth and risk in 2013.


245



It was found that the distances identified using Multidimensional Scaling Analysis of the QISMUT countries and their disparities have a linear relationship. According to the results of Euclidean distance model for 2014, it was found that the estimated distances were consistent with the actual values as the estimated values were similar to the ones in the distance matrix obtained from data.



246

According to the Figure 6, 53 out of 56 banks included in the analysis have financial similarities (accumulated around Euclid) as per the Euclidean distance model which shows the relations between units (QISMUT Islamic banks) on a 2-dimensional space. According to the Euclidean distance model for 2014, 3 banks (I12: PT Bank Syariah Panin, M6: Bank Pembangunan Malaysia Berhad and S2: Alinma Bank) have financial disparities when compared to the other Islamic banks. Therefore, Islamic banks operating in Qatar, UAE and Turkey have similar properties in terms of profitability, growth and risk in 2013.

As it is shown in the Euclidean distance graphs, Islamic banks operating in QISMUT countries are quite similar in terms of their financial ratios collected for 2012 and 2013, thus, clustered very close to each other in the graph. Only 3 of the banks were distant from this cluster. In 2014, on the other hand, distances between banks clustered are slightly expanded when compared to the other years.

5. CONCLUSION

Islamic banking activities, one of the building blocks of Islamic finance, have recently been developing in the global arena. Growing its market share in the global economy, Islamic finance is led by the countries called QISMUT in the recent years. Consisting of Qatar, Indonesia, Saudi Arabia, Malaysia, UAE and Turkey; QISMUT is anticipated to drive the Islamic capital which rapidly grows its market share in the global finance system. Efficient operability of Islamic banks is important for the healthy development of many economies when this substantial market share is taken into consideration.

This study identified the financial similarities and disparities between Islamic banks using cluster analysis and multidimensional scaling analysis. Financial similarity and disparity levels of Islamic banks operating in the QISMUT countries were duly measured. Financial similarities and disparities of Islamic banks were then separately evaluated for a period between 2012 and 2014 using profitability, growth, risk and DuPont ratios.

According to the cluster analysis results for 2012, Islamic banks operating in Qatar, Saudi Arabia, and Turkey have similar properties in terms of profitability, growth and risk. For the same year, 53 out of 56 banks included in the analysis have financial similarities (accumulated around Euclid) as per the Euclidean distance model and Multidimensional Scaling Analysis. According to the Euclidean distance model for 2012, 3 banks (I12: PT Bank Syariah Panin, M6: Bank Pembangunan Malaysia Berhad and S2: Alinma Bank) have financial disparities when compared to the other Islamic banks.

As it was the case for 2012, PT Bank Syariah Panin of Indonesia was financially different when compared to the other banks included in the sample also in 2013. Islamic banks operating in Qatar, Saudi Arabia, UAE and Turkey have similar properties in terms of profitability, growth and risk in 2013. 53 out of 56 banks included in the analysis have financial similarities (accumulated around Euclid) as per the Euclidean distance model which shows the relations between QISMUT Islamic banks on a 2-dimensional space. According to the Euclidean distance model for 2013, 3 banks (I12: PT Bank Syariah Panin, M6: Bank Pembangunan Malaysia Berhad, and S2: Alinma Bank) have financial disparities when compared to the other Islamic banks. Therefore, Islamic banks operating in Qatar, UAE and Turkey have similar properties in terms of profitability, growth and risk in 2013.


247

According to the Cluster analysis results for 2014, Islamic banks operating in Qatar and Saudi Arabia have similar properties in terms of profitability, growth and risk. For the same year, 53 out of 56 banks included in the analysis have financial similarities (accumulated around Euclid) as per the Euclidean distance model. According to the Euclidean distance model for 2014, 3 banks (I12: PT Bank Syariah Panin, M6: Bank Pembangunan Malaysia Berhad, and S2: Alinma Bank) have financial disparities when compared to the other Islamic banks. Therefore, Islamic banks operating in Qatar, UAE and Turkey have similar properties in terms of profitability, growth and risk in 2013.

As it is shown in the Euclidean distance graphs, Islamic banks operating in QISMUT countries are quite similar in terms of their financial ratios collected for 2012 and 2013, thus, clustered very close to each other in the graph. Only 3 of the banks were distant from this cluster. In 2014, on the other hand, distances between banks clustered are slightly expanded when compared to the other years.

In line with the findings of this study, it would be fair to say that QISMUT countries which will make their mark in the future of Islamic finance and banking have similarities in terms of financial ratios.

This study is an original research measuring the similarities and disparities of Islamic banks operating in QISMUT countries which are expected to lead the future of Islamic banking and finance. In this respect, this study will make a significant contribution to the literature. Future research may focus on extended numbers of Islamic banks operating in QISMUT countries and may diversify the research periods and variables.

REFERENCES

Aldohni, A. K. (2015) the Quest for a Better Legal and Regulatory Framework for Islamic Banking, Ecclesiastical Law Journal / Volume 17 / Issue 01 / pp 15 – 35.

Amel Belanès, A., Ftiti, Z. and Regaïeg, R. (2015) What can we learn about Islamic banks efficiency under the subprime crisis? Evidence from GCC Region, Pacific-Basin Finance Journal 33 (2015) 81–92

Ariff, M. (2014) Whither Islamic Banking? The World Economy, Volume 37, Issue 6, pages 733–746.

Bellalah, M. and Ellouz S. (2004). Islamic finance, interest rates and Islamic banking: A Survey of the literature. Finance India: special issue, 18, 533-546.

Burn, D.H. and Goel, N.K. (2000) The Formation of Groups for Regional Flood Frequency Analysis. Hydrological Sciences Journal, 45(1), 97–112

Dardaci, N. and Boitan, A.I. (2009) A Cluster Analysis Approach for Banks’ Risk Profile: The Romanian Evidence, European Research Studies, Volume XII, Issue (1) 2009

Ernst and Young; World Islamic Banking Competitiveness Report 2013–14


248

Grzegorz H. and Dawid Z. O. (2009). Strategic groups and banks’ performance, Financial Theory and Practice, 33 (2), pp. 153-186.

Hair, J. F., Anderson, R.E. and Tatham, RL. (1998). Multivariate Data Analysis, New Jersey: Prentice Hall Inc.

Hair, J. F., Black, W., Anderson, R., Babin, B. R. and Tahtam, R. L. (2006). Multivariate Data Analysis With Readings. Mcmillan Book Company: London.

Han, J. and Kamber, M. (2006) Bibliographic Notes for Chapter 2Data Preprocessing, http://www.academia.edu/2238328/Data_Mining_Concepts_and_Techniques_2nd_edition_

Hand, D., Mannila H. and Smyth P. (2001) Data Mining. Massachusetts Instittute of Technology, USA

Izenman, A. (2008) Modern Multivariate Statistical Techniques, USA, Springer.

Knotek, P. (2014) Banking Sectors In Emu – Cluster Analysis, European Scientific Journal December 2014 edition vol.10, No.34.

Kowal, R., Yeleyko, Y.I. and Kharkhalis, V.I. (2014) Application of Cluster Analysis to Assess the Regional Development of Foreign Banking in Ukraine, http://miscellanea.ujk.edu.pl/data/Oferta/Pliki/490_18-Kowal.pdf

Lackmann B.G. (2014) The Six Key Countries Driving Global Islamic Finance Growth, Nomura Journal of Capital Markets, Vol. 6, No:2.

Lin, G.F. and Chen, L.H. (2006) Identification of Homogeneous Regions for Regional Frequency Analysis using the Self-Organizing Map. Journal of Hydrology, 324.

Liu, H. and Yu, X. (2009) “Application Research of k-means Clustering Algorithm in Image Retrieval System” Proceedings of the Second Symposium International Computer Science and Computational Technology, Huangshan.

Lucotte, Y. (2015) Euro area banking fragmentation in the aftermath of the crisis: a cluster analysis, Applied Economics Letters, 2014, http://dx.doi.org/10.1080/13504851.2014.997918

Machado J. and Mata M. (2013) Analysis of Stock Market Indices through Multidimensional Scaling. The Scientific World Journal. doi: 10.1155/2013/594587

http://miscellanea.ujk.edu.pl/data/Oferta/Pliki/490_18-Kowal.pdf

http://dx.doi.org/10.1080/13504851.2014.997918


249

Machado J. and Mata M. (2015) Analysis of World Economic Variables Using Multidimensional Scaling. PLoS ONE 10(3): e0121277. doi:10.1371/journal.pone.0121277

Mallin, C., Faraga, H. and Yonga, K. (2014) Corporate social responsibility and financial performance in Islamic banks, Journal of Economic Behavior & Organization 103, S21–S38.

Mar Molinero, C. and Serrano Cinca, C. (2001) Bank failure: a multidimensional scaling approach, European Journal of Finance, Vol 7, No 2, June, pp. 165-183, Ed Routledge.

Mghaieth, A. and Mehdi, I.K.E. (2014) The determinants of cost/profit efficiency of Islamic banks before, during and after the crisis of 2007-2008 using SFA approach, IPAG working papers, http://www.ipag.fr/fr/accueil/la-recherche/publications-WP.html

Nazim, A. and Bellens, J. (2014) World Islamic Banking Competitiveness Report 2013–14, www.ey.com

Neil, T. H. (2002). Applied Multivariate Analysis, Secaucus, NJ, USA: Springer-Verlag New York.

Rao, A. and Srivinas, V.V. (2006). Regionalization of Watersheds by Fuzzy Cluster Analysis. Journal of Hydrology, 318, 57-79.

Smola, E. and Mirakhor, A. (2010), The Global Financial Crisis and Its Implications for the Islamic Financial Industry, International Journal of Islamic and Middle Eastern Finance and Management, 3, (4), 372-385.

Sorensen, C. K. and Gutierrez, J.M.P. (2006) Euro Area Banking Sector Integration Using Hierarchical Cluster Analysis Techniques, European Central Bank, WORKING PAPER SERIES, NO 627.

Vagizova, V. and Lurie, Ksenia, Ivasiv, I. (2014) Clustering of Russian banks: business models of interaction of the banking sector and the real economy, Problems and Perspectives in Management, Volume 12, Issue 1, 2014

http://www.ipag.fr/fr/accueil/la-recherche/publications-WP.html

http://www.ipag.fr/fr/accueil/la-recherche/publications-WP.html

http://www.ey.com/

financial differences and similarities of islamic … · similarities between banks in terms of...

Documents